Med Educ. 2026 Aug 23. doi: 10.1111/medu.70290. Online ahead of print.
ABSTRACT
BACKGROUND: Errors are an inevitable part of clinical learning, yet procedural training commonly emphasises error avoidance. This study examined the effect of deliberate replication of predefined procedural errors (Error Replication; ER), compared with error avoidance training (EAT), on post-test performance, retention and transfer in surgical simulation.
METHODS: In this randomised controlled trial, 70 medical students with no prior simulation-based osteosynthesis experience completed a pre-test of simulated dynamic hip screw (DHS) fixation before 60 min of ER or EAT training. ER participants deliberately reproduced predefined, clinically relevant errors, whereas EAT participants used simulator feedback to improve performance and minimise errors across successive iterations. Participants then completed an immediate DHS post-test, a DHS retention test after 6-10 days and a transfer test involving antegrade femoral nailing. Transfer-test performance was the primary outcome. Secondary outcomes included performance trajectories across pre-test, post-test and retention test; completed DHS iterations; self-perceived dexterity and spatial awareness; and perceived training efficacy.
RESULTS: Complete data were available for 68 participants. Performance improved substantially over time in both groups (p < 0.001, η2 p = 0.606), with no significant group-by-time interaction (p = 0.44). Transfer scores were 52.6 (SD 24.7) for EAT and 51.0 (SD 33.9) for ER (p = 0.82; d = 0.06, 95% CI – 0.42 to 0.53). No statistically significant differences were detected in the remaining secondary outcomes.
CONCLUSIONS: Participants improved following both instructional approaches, although those in the ER group spent most of the training period deliberately reproducing predefined errors. No statistically significant differences between ER and EAT were detected in post-test performance, retention or transfer. These findings suggest that structured exposure to clinically relevant procedural errors warrants further study as a distinct instructional modality.
PMID:42633790 | DOI:10.1111/medu.70290
